JOB AUTOMATION AND ITS EFFECT ON EMPLOYMENT

This project mainly involved performing analysis through data visualization using Tableau. As automation grows as a trend, many manual jobs are being replaced. For example, many CVS stores have replaced cashiers with self-service checkout machines. Through this research, we wanted to see whether likelihood of automation for jobs correlates with actual changes in job markets, such as mean wage, number of jobs, etc across different industries in United States of America.

We collected the data about employment statistics from the United States Bureau of Labor Statistics website from the years 2012 – 2016 covering total employment, occupation, hourly and annual wages etc. We added a variable called automation probability to these data based on occupation code. The automation probability data was retrieved from this dataset (Click Here).

We did some statistical modeling and machine learning to gain more insights into the data. Using ANOVA in R, we compared the change in mean wages to change in total employment over the years 2012-2016 and found that Employment increased at a higher rate over the five years span than wages. Through neural net, we found that location quotient is most significant factor for change in number of jobs from 2012-2016, implying that if a job is more common in area, automation will have lesser impact on job numbers.

Through our research, we found that average automation probability is pretty same across all the states in USA, for exception like District of Columbia. Jobs like surgeons which are less likely to be automated soon have a higher mean wages. Even though jobs with higher probability of automation are increasing every year, there is a dip in percentage increase of their numbers and might see a downfall soon.

Below is the video of project report for your reference. This was a team project and my team members were Nakul Kaura, Evan DeCastros and Yicong Hu.